Blog post

Long term real interest rates fell below zero in all euro area countries

The 10-year real government bond yield, which is the nominal yield deflated by expected inflation, has fallen below zero in Italy and Greece, boosted

Publishing date
08 October 2019
Zsolt Darvas

While nominal interest rates of all euro-area countries converged to the German rates during the first decade of the euro, large differences developed with the emergence of the euro crisis after 2009: interest rates in southern European countries and Ireland increased significantly in 2010-2013, while the rate in Germany and other top-rated countries decreased. The differences in nominal interest rates were translated to even larger differences in real interest rates, that it, nominal rates adjusted by expected inflation. Expected inflation in Germany was (and still is) higher than in southern Europe, thereby driving a large gap in real interest rates.

Low real interest rates greatly help borrowers, including the public sector, and might also boost investment and growth. At the same time, they are bad for savers. Conversely, high real government bond yields could undermine fiscal sustainability and pull private sector interest rates up, making corporate and household borrowing more costly, which weakens economic recovery. Therefore, divergent interest rate developments in the euro area is a concern and it could also undermine the transmission of ECB’s monetary policy.

The good news is that the interest rate divergence narrowed significantly in recent months. In this post I calculate the 10-year real government bond yields for 27 EU countries in comparison with Japan and the United States. I found that the real rates have fallen below zero in every euro-area country, and with the exception of Romania, in all EU countries too.

How to measure the long-term real interest rate?

An important question is how to measure inflationary expectations 10 years ahead? Some analysts use the actual inflation rate, such as the inflation from last year to the current year, and deduct this from the nominal interest rate. But such a simplistic assumption is inadequate, because inflation in a particular year could be affected by unusual developments, like large swings in energy prices, which might not be expected in the future. Market-based inflationary expectations, such as inflation indexed bond yields and swaps, are useful indicators, but these are available only for a few countries. There are surveys of long-term inflationary expectations, but those are sporadic and available only for a few countries.

I have therefore decided to use inflation forecasts from the IMF World Economic Outlook, which are available five years ahead starting from the forecast made in April 2008 (before 2008, the IMF published forecasts only two years ahead). I assume that the inflation forecast for the fifth year ahead will prevail in the subsequent five years too, and thereby I calculate inflation projections ten years ahead.

Certainly, IMF forecasts (as well as all other forecasts) are subject to large errors. But this is not an issue from perspective of calculating real interest rates, because economic decisions depend on expectations. For example, ten-year forward-looking economic decisions made in 2019 depend on expected inflation in 2019-2029 and cannot depend on actual inflation in 2019-2029 (because that will be known only in 2029). If IMF forecasts represent expectations, then such forecasts are useful in calculating real interest rates, even if actual inflation in 2019-2029 will be different from the forecast.

IMF forecasts are typically published in April and October of each year, but I would like to calculate real interest rates for each month. I have therefore linearly interpolated the April and October ten-year ahead inflation forecasts. In most years this is probably a harmless assumption, given that long-term inflation forecasts hardly change from one forecasting round to the next. The exception was 2009, when the global financial crisis induced some larger swings even in long-term inflation forecasts.

Euro-area long-term real interest rates

The first euro-area country that faced a negative ten-year real interest rate on newly issued government bonds was (unsurprisingly) Germany in September 2011 (see the charts below). By January 2015, thirteen eurozone countries joined this club – the exceptions were Cyprus, Greece, Italy, Portugal and Spain. (Estonia is not considered due to the lack of ten-year government bond yield data, resulting from the very low level of public debt).

Real interest rates continued to fall throughout the euro-area recent years, with the German ten-year real rate falling to a stunning value of -2.7% per year in September 2019. There are nine other euro-area countries with a real rate at or below -2%. Spain joined the distinctive club of countries with negative ten-year government bond yield in July 2016, Cyprus in March 2019, Portugal in May of the same year, Italy in August and Greece in September. In the latter two countries, positive market assessment of the new governments played a major role in recent interest rate falls.

While low negative rates are great news for government treasuries*, they are bad news for government bond investors. For example, the -2.7% German ten-year real yield means that if someone invest her/his saving into a ten-year German government bond, the real value of that saving is expected to decline by 2.7% in each year. By the end of the tenth year, the real value of the saving falls by almost a quarter, that is, a quarter less goods and services can be purchased in ten years from the result of the saving than the amount of goods and services the same saving could purchase now.

The opposite effect applies to the government. The government can decide to borrow now at a -2.7 annual rate for 10 years, spend the proceeds of the borrowing now, and repay this loan from taxes collected in 10 years’ time. Or it can decide not to borrow now, but collect the taxes in 10 years’ time and spend the money then. The tax burden is exactly the same in both cases, but in the first case the government can purchase a quarter more goods and services now than in the second case in 10-years’ time. Clearly, unless there are concerns about fiscal sustainability, it is high time for the government to borrow and invest in the future – see my concluding thoughts on this issue at the end of my post.

The recent falls in long-term interest rates are most likely driven by a change in market interest rate expectations, whereby monetary tightening is not expected in the foreseeable future. In Figure 2 I reproduced and updated the shocking chart from a report by Grégory Claeys, Maria Demertzis and Francesco Papadia, which shows how market expectations of the short-term euro interest rate changed from mid-April to mid-August 2019, that is, a month before the 12 September monetary policy meeting of the ECB Governing Council. We updated the chart with early October data. While in April markets expected that the short-term euro rate will remain negative till 2025 and by 2030 it will be somewhat over 1%, the August expectations foresaw a negative rate till 2030, a major fall in expectations just over four months. There was only a slight correction from August to October. Moreover, the expected rate for 2049 was reduced from 1.17% in April to -0.1% by August, which just moved up slightly by October (to +0.1%).

The reasons for this huge shift in expectations (even for 2049) during the four months between April and August is unknown; I have some doubts whether the worsening of the near-term global outlook would justify such a big fall in expected interest rates in 30 years’ time. The fall in inflation expectations cannot explain either, because the market-based inflationary expectations of the euro area over the next ten years have fallen just by 0.25 percentage points from 1.23% in April to 0.98% in September, which is much smaller than the fall in the expected average short-term interest rate over the next ten years. For interest rate expectations further ahead, such as for 2049, not the average inflation over the next years, but the long-term inflationary expectations matter: the expected annual euro-area inflation in ten years’ time fell even less from 1.40% in April to 1.34% in September suggesting that the fall in expected short-term interest rates in the decades to come almost entirely reflects an expected fall in the real interest rate too.

Yet whatever the reason, if markets have substantially lowered their expectation for short-term interest rate in the next three decades, it is not surprising that 10-year government bond yields also fall further. Such changes in interest rate expectations, combined with the expected decline of the stock of government bonds as a share of GDP, induced investors to accept even lower negative yields on government bonds.

Therefore, in my view the change in interest rate expectations, and not the September monetary stimulus of the ECB, have led to a further fall in government bond yields. In any case the announced resuming of asset purchases from November at €20 billion per month is just one-fourth of the amount purchased in 2016. Yet real interest rates fell to much lower values now than in 2016.

Non-euro area countries

The large fall in real interest rates extended to non-euro area countries too. Denmark, a country that pegs its currency to the euro, has the second lowest ten-year real interest rates in the EU at -2.4% in September. Sweden is not far with a rate of -2.1%, nor Bulgaria with a -1.9% rate. The UK’s real rate at -1.4% in September 2019 is also quite low. There were also large real interest rate falls in the four central European countries with a floating exchange rates, the Czech Republic, Hungary, Poland and Romania, leading to negative ten-year government bond yields, with the sole exception of Romania.

Interestingly, Japan, a country that’s had low interest rates for more than two decades now, has a real long-term rate of -1.5%, which is just in the middle of the range of EU countries. 16 EU countries have an even lower rate.

In the US, the real rate increased from negative values in 2016 to about 1% in 2018, due to the monetary tightening of the Federal Reserve, but more recently it fell to -0.5% by September 2019. There are only three EU countries (Romania, Poland and Greece) that have a larger rate. The ‘exorbitant privilege’ arising from the central role of the US dollar in the international monetary system has not prevented 24 EU countries (or 25, if Estonia could be considered too) from having a lower real interest rate on ten-year government bonds than the US, even in some small countries whose currency has practically no international role, like Bulgaria, Croatia, Czech Republic and Hungary.

Policy implications

The negative real government bond yields, which are well below the expected real economic growth rates, create a lot of additional fiscal space, as it was prominently highlighted by Olivier Blanchard.

The fiscal space created by low interest rates should be used wisely. In our memo to the new EU commissioner for economic affairs, jointly written with Thomas Wieser and Stavros Zenios, we recommended channelling the interest savings to investments, such as the transition to a carbon-neutral economy, rail and road, research and development, and digital infrastructure. There are pressing public investment needs in many countries and the negative real interest rate environment presents the perfect opportunity to finance such needs. Moreover, in the current environment of deteriorating economic outlook and low inflation, at a time when the European Central Bank has limited scope for additional monetary easing, a temporary discretionary fiscal boost is the right policy answer.

However, while real government bond yields fell everywhere, differentiation is necessary because the highly-indebted countries should not repeat past mistakes of pretending that they can spend their way out of the debt sustainability trap.

And policymakers should recognise the risks of a possible medium-term reversal of interest-rate developments, which has two main implications. First, additional spending should focus on categories that are easier to reverse, e.g. it is easier to reverse an investment stimulus than increased entitlements. Second, governments should ‘lock-in’ as much low interest rate debt as possible, for example by rolling over maturing short-term debt with long-term debt.

The IMF has made available its historical five-year ahead forecasts starting from the spring forecast of 1990, for three indicators: GDP growth, inflation and current account balance. We use this dataset to update our long-term real interest rate calculations starting from 1990, whenever 10-year government bond yield data is available. We also use data from the most recent October 2019 World Economic Outlook dataset to update inflation forecasts. The interest rate data for October 2019 corresponds to the average of the first 25 days of the month.


* I emphasise that I consider the real rate on new borrowing, while government bonds issued earlier have their own –typically larger– interest rates to pay. But the longer the period of negative rates lasts, the lower the total interest burden becomes.

About the authors

  • Zsolt Darvas

    Zsolt Darvas is a Senior Fellow at Bruegel and part-time Senior Research Fellow at the Corvinus University of Budapest. He joined Bruegel in 2008 as a Visiting Fellow, and became a Research Fellow in 2009 and a Senior Fellow in 2013.

    From 2005 to 2008, he was the Research Advisor of the Argenta Financial Research Group in Budapest. Before that, he worked at the research unit of the Central Bank of Hungary (1994-2005) where he served as Deputy Head.

    Zsolt holds a Ph.D. in Economics from Corvinus University of Budapest where he teaches courses in Econometrics but also at other institutions since 1994. His research interests include macroeconomics, international economics, central banking and time series analysis.

Related content